@olivierlambert Looks possible, since I've clicked forget this particular slave/host in XOA, how would I go around adding it back there?

The slave is missing from the masters pool list.

Hi there, just recently did a bunch of changes to my internet infrastructure and lan, I've got everything working again except one of my hosts.

It's complaining that the master node is unreachable, this is due to an IP change, I've set a static one within OPNSense so that shouldn't be a problem in the future.

I can't seem to find a way to change the IP for this Host in XOA within the pool manager. I've tried forgetting it, I've tried removing it from the pool via keyboard and vga... says it isn't in one. Xapi is throwing errors and is unresponsive.

How should I go around solving this kerfuffle?

Any help is greatly appreciated, thanks. - MrXeon.

@olivierlambert

I've solved the issue a little while back, it was definitely user error.

OPNsense expects the theoretical "LAN" to be highest priority when initializing and the theoretical "WAN" to be second..

Here a screenshot of the working method within XO...

@MrXeon Well I'll correct myself regarding a few things.

Vlans aren't necessary and won't work in this scenario also there's nothing wrong with the default Pool wide network interfaces, I was simply misconfiguring them for this specific VM, OPNsense expects xn0 to be the "LAN" and xn1 to be the "WAN" thus in my case being misconfigured in the opposite order, I would assume it was trying to bind the "WAN" as "LAN" which made the services hosted on that Network interfaces via the XCP-ng host unusable...

Simply swaping these around in the order OPNsense expects fixed the issue

The only problem I've come across so far is that OPNsense can't ping the internet, where's other VM's can on that interface... so I would assume that's actually an issue with my ISP's Router behind my AP so gonna do some further testing and figure that out.

@olivierlambert Another update, perhaps it was user error... I'm not sure though..

Initially I was trying to pass the default Pool wide network interfaces, this would just end up making the host Inaccessible from XO.

When creating two new new network interfaces assigned to eth0 and eth1, under my pools network tab and attaching those to my OPNsense VM, I find that it no longer causes issues and runs as one would expect.

Is this the proper setup and I've simply made a mistake or is there something goofy going on with my system?

I will note that I defined two separate vlans for those newly created networks one labeled as WAN and the other labeled as LAN respectively...

My goal is to utilize the OPNsense VM to completely take over networking in my environment, providing internet access to other VM's and pyshcial equipment via switches/access points..

Is this the proper configuration to achieve this desired task?

@olivierlambert It might be an IP conflict.

When removing both default pool wide network interfaces (My system has dual nics) OPNsense starts as it should.

Is the VM by chance being assigned the same IP as my XCP-ng host thus knocking out all services, how should I go around fixing this, is it a bug or user error?

@olivierlambert My XO and NFS server share their own IP Address (Same Machine), my XCP-ng Host definitely has it's own IP Address.

Not sure about the VM itself, I've added both pool wide network interfaces to it's configuration. (Eth0, Eth1) dual nics.

If that sounds correct, how should I go around checking the logs, /var/log I would assume?

Hi there, working with a clean fully functional install of XCP-ng on my Host Machine, I've created a Debian 11 VM successfully..

when attempting to setup a OPNsense (PFsense) fork, using a reasonable amount of resources, I notice that during first boot/installation the VM causes the stack to become unresponsive including SSH connection to the host, occasionally I'll regain connection but it'll then drop again.

Any ideas why this is happening?

Regards,

MrXeon.

@Andrew I’ve started to experiment with Duet + OpenCore chain loading but I'm getting the same results…

I’ve tested it now on two different Legacy Machines, so either I’m setting it up wrong or Duet is the issue all around the baord… why could Duet potentially not be happy?

It’s not my target system, it’s gotta be something to do with the file structure or the DuetPkg itself, I've also checked for Debug Logs from OpenCore, seems it ain't booting at all, so it's definitely an issue early on.

@Andrew where did you find the CloverLegacyInstaller?

@Andrew When you get a chance could you explain to me the method you used for flashing Clover to a USB Key?

I would appreciate that since for the life of me I can't create a working Legacy Clover Key haha, I'm not too sure why my methods of flashing didn't work but it sounds to me based on it being messy that you've manually installed it rather then used a automated utility.

Quite curious to know what the steps to achieve that on a Windows or Linux Machine would be.

@Andrew wow you honestly got further then me and I've been trying for a while haha, way to go!

What utility did you use to flash the Clover image to your USB stick?

Here's my most recent Clover Attempt Story, I went through the trouble of running Windows on my target machine because a few forum users mentioned that it works best if you flash it on your target system, to be honest I'm not sure how that would really make a difference but I tried it anyway.

First off downloaded a Windows 10 Home ISO from Microsoft using a Browser with a spoofed User Agent, Flashed it to a Sata SSD via a USB 3.0 Encloser/Adapter in around 3 Minutes using the Freeware utility Win 2 Go, and finally connected that SSD to my Target Systems SATA Port.

That method allowed me to skip the Windows Media Creation Tool and The Windows 10 Installer, I was booting straight to Windows with a few basic system setup steps.

For anyone trying to reattempt that, I would suggest you try self customizing the Windows 10 ISO to not include all the extra bloatware and unnecessary setup features… it’ll save a little extra time. (However I didn’t end up doing that part)…

So after one blue screen of death… lets just say Windows got a little overwhelmed, I was able to Run the Boot Disk Utility and flash Clover to a USB key.

After ejecting the flash drive and waiting for Windows 10 to shutdown, I disconnected my SSD and Inserted the Clover Key… unfortunately it still Boots to a 6 with a Blinking White Cursor and stays there.

I'll add that I was trying to boot into a very old version of Clover v4961 using BDU (v2.1.2018.023b)... however I've tried just flashing the latest version via Rufus with similar end results.

Unless it's being picky about my system in particular I would guess you did something different that I'm missing during the early steps.

@Andrew I'm planing to use the TrueNasCore Instance as a file host unrelated to my XCP-ng infrastructure, however I might consider using it as a ISO SR as well, otherwise using zfs on dom0 is a great idea, just isn't preferable for my use case.

Unfortunately I don't have access to a 4th sata drive, I've got 3 sata bays going to a raid controller/backplane, 2 USB 2.0 Ports, 1 USB 2.0 Header, and 1 PCie 2.0 Slot.

I didn't consider sata type b ssd's when planing the setup, now I kinda wish I did, I could switch gears if Clover doesn't work out. Do such cards provide power and data transfer through the PCie lane?

Regarding Clover, I've been experimenting with it for a bit before I tried moving the Boot directory to a Separate flash drive... When using the latest version and flashing it via Rufus on a separate Windows 10 Machine, I couldn't get the Clover key to boot on my target machine, it kept leaving me at a white blinking cursor which is usually a sign that I either incorrectly flashed it or it's corrupted in some way maybe even not compatible with my system.

This seems to be an issue with the latest versions of Clover when attempting to boot on Legacy BIOS systems acording to one Clover user.

So I attempted to use a much older version that supposedly works better on older systems, this time I actually reached a number 6 with a white blinking cursor, which is a sign that it's compatible however it never reached the Clover Menu which means I incorrectly flashed it.

I've been told that the recommended flashing utility is BDU "Boot Disk Utility" created to make Clover installations easier, however do be warned that many copies of BDU and other Clover Flashing utilities floating around the Internet unfortunately contain Malware, so it's best to attempt to find clean copies amongst popular bios modding community members.

I'm not sure if the initial developer intended for BDU to be malicious or if it just kinda became abandoned ware and malicious actors have taken it over, a good example being the AIO Boot Tutorial you shared, Who knows why but the creator actually infested their utility with Malware and once people complained they took it down silently.

I was able to get my hands on a working safe version of BDU from a forum user so I'm currently testing that, I've been told I'll have better luck if I flash it on the Target system itself, so I'm setting up a Windows 2 Go live bootable as I write this, So wish me Luck.

It would be nice to know if someone with a similar but slightly different setup runs into the same issues or not with Clover, so if you get the chance please do let me know how it works out for ya.

If I can't get Clover working then I think my best bet might be to start looking into those Type B Sata SSD Adapters you mentioned, could you recommend any particular one from experience?

@Andrew That sounds like a reasonable plan, but won't TrueNasCore need complete access to all 3 sata drives at least in order to do software raid?

I don't think it'll like having a partial compacity of the drive.

Pretty certain I have to do sata pass-through.

@Andrew Did some further testing as you suggested and your definitely bang on, it's a migration/mounting problem most likely caused by moving Grub and the Kernel to my Flash Drive.

When booting from a SATA SSD I've noticed some interesting changes which I've noted below.

XCP-ng loading screen.

ACPI Error during boot (could you help me troubleshoot this?)

Text is white and not bold with colorful font like the previous setup.

Configuration Screen has changed Resolution/aspect ratio, now it's 4:3 while before it was 16:9.

Networking works out of the gate.

Dual Xeons are detected

My first question is regarding the slight changes... why?

Do you have any suggestions on how I might be able to work around this? Are there any configuration files I can change to correct the mounts or am I stuck with having to reconfigure my drive setup?

The only other thing I can think of is to continue attempting to get Clover to boot, if I can do that I can keep everything on the NVMe but it's questionable if that's worth my time.

@Andrew I just reinstalled for the third time.

Fresh downloaded ISO, passed signature check, flashed via dd on Arch Linux to a 3.1 USB Flash Drive fresh out of the package.

XCP-ng Installer Live Booted successfully and passed it's medium check.

Installed to my NVMe and then Booted into a Debian Based Live os (Sparky Linux) and moved my boot directory over to a flash drive and installed Grub to it's MBR...

Shutdown, powered back up, XCP-ng networking wasn't working so I did a emergency reset (This seems to happen every fresh install) and then did a "yum upgrade" before trying to run xl dmesg again.

Unfortunately it gave the same error

@Andrew I'm not using Clover it didn't work for me, I simply moved the boot directory to a flashdrive after installing XCP-ng to my NVMe.

My Supermicro Server only has 3 sata drives per node, which I want to dedicate to my Software Raid/File Server which will be running as a VM (TrueNasCore) hopefully under XCP-ng.

Thus why I'm using my PCie 2.0 slot for NVMe, the only other option would be to use the 3 drive bays as a hardware raid 5 of some smaller sata SSD's and use that for my SR and Host, then repurposing the PCie 2.0 slot for some sorta controller or high speed nic to communicate with a separate NAS Box of sorts.

Least that's the only work around I've thought of, otherwise I would have to use USB 2.0 for the Host and Guest Systems which isn't very ideal.

@Andrew OH it's important to note that I've moved the boot directory over to a flash drive prior to first boot (changes applied to fstab)

The reason for this is because I've got a BIOS Based Server which doesnt support NVMe, however I wanted to boot from a NVMe or at least keep the Host and Guest Filesystems stored there, so moving the Boot directory (Grub and the Kernel) over to a flash drive seemed like the best work around since I wasn't able to get Clover or similar bootloaders with "UEFI emulation" support to boot on my system.

The original boot partition is still intact just being ignored, the original boot directory was renamed to boot_old for safe keeping and those files were copied over to a ext4 mbr based flash drive mounted to the new "boot" directory.

I don't think that should cause these problems, but just in case that caused something unpredictable I thought I would mention it.

@Andrew I'm running it directly on the local command shell, My XCP-ng disk image passed a signature verification and passed the media check during installation.

First install had a corrupted backup partition however (which I ended up deleting and repurposing out of inexperience thus this is my second fresh installation).

This time the backup partition wasn't corrupted.

What would you suggest I do?

@Andrew Gave that a shot and got the following error.

["xl dmesg" output from XCP-ng Host]

Xencall: error: could nor obtain handle on privileged command interface: No such file or directory.

Libxl: error: libel.c:libxl_ctx_alloc: cannot open libxc handle: no such file or directory.

Cannot init xl context